ABSTRACT
An index defined for a country can be applied also to its regions. This is particularly relevant for the largest European countries, which have many regions and where substantial differences among the regions are likely to be present. When a region forms a substantial part of the country and is represented in the survey by a substantial part of the national sample, we simply apply the methods and procedures devised for a country. One important exception in this is that the national standard (e.g., a given percentage of the national median equivalised household income (eHI) used as the poverty threshold) is applied instead of a standard specific to the region (based on the median eHI for the region). When a region is represented in the survey by a relatively small sample, for which the estimates would be associated with large standard errors, we seek some improvements by means of small-area methods elaborated in the next section. In their established form, they are relevant in the European Union Statistics on Income and Living Conditions (EU-SILC) up to 2010 only for Spain and France, for which, respectively, 19 and 22 regions are defined in the data. For other populous countries, no regions are defined for the UK and Germany and only five regions for Italy and six for Poland. Czech Republic has eight regions; the other countries have at most four regions identified in the data. Several of the smallest countries form a region on their own (Cyprus, Estonia, Iceland, Luxembourg, Malta and Slovenia).
